Do we ever learn from our past mistakes?
For many years we believed that technology was an inevitable force for good. It would give us instant access to a near infinite amount of information and allow us to easily and instantly connect with nearly anyone on earth. What could go wrong? The answer is that there are always unintended consequences, and we didn’t anticipate that the same technology would also unleash dangerous dystopian forces that are now eroding our politics and our social structures.
Now the same sort of breathless optimism about technology is infecting medicine and healthcare. Apple, Google, and many other companies are racing to bring their technologies to healthcare, which they view as a vast untapped frontier. These companies are forging ahead without any evidence that they’ve learned from their past mistakes, and the rest of us seem to be all too willing to allow ourselves to be fooled again.
The Apple Heart Study, published this week in the New England Journal of Medicine, is the most significant instance to date of big technology’s invasion of the healthcare domain. This enormous study looked at the role of the Apple Watch in detecting atrial fibrillation (AF), which is the most common heart rhythm disorder and a significant cause of stroke and declining quality of life in older people.
Over the past few decades a series of important studies led to remarkable advances in the understanding and treatment of AF. The Apple Heart Study was completely different from all previous AF studies. It was conducted online and involved no physical contact between the study subjects and the researchers. Instead, this “virtual” study recruited subjects by inviting Apple Watch users to download an app and enroll themselves in the study.
The most remarkable and, at first, impressive fact about the study is its size. While traditional studies struggle, at great expense, to recruit a few hundred or a few thousand participants, the Apple Heart Study was able to recruit an astonishing 419,297 participants in a few short months. Nothing like that has ever been seen in medical research.
So what’s the problem? After all, if you can study hundreds of thousands of people for less money and in a shorter time why would that not be a good thing? But a close look at those participants, and a careful scrutiny of the findings of the trial, make clear that the trial does not represent anything resembling a genuine advance in medical research, and it raises troubling questions about the role of technology, and big tech companies, in healthcare and medical research.
As a general rule if you want to study a disease you study people who actually have the disease or, at least, are at high risk for developing the disease. When it comes to AF this means an elderly population, because the vast majority of people with AF are over 65 and younger people rarely develop it. But more than half the patients in the study were under 40 years of age and only 6% (24,626) were over 65, the age group most at risk for AF.
As it turns out, the Apple Watch detected an irregular pulse in only 341 (0.16%) of the 219,179 people under 40 and most of these did not turn out to have AF upon further investigation. Even in the 65 or older group only 3.2% had an abnormal rhythm detected and, once again, only a small portion of these turned out to actually have AF. Even these percentages are suspect, since a huge number of people who enrolled in the study simply stopped participating.
In other words, nearly half a million people were required to identify a few hundred people with AF. (It should further be noted that we have no idea whether the patients identified in this group will even benefit from their diagnosis, since they may differ in important ways from AF patients identified by traditional means.)
These numbers suggest that the Apple Heart Study was a colossal waste of money and resources. No actionable knowledge was gained. But the study does highlight the dangers of blindly or indiscriminately adopting technology in healthcare. In an accompanying editorial two editors of the New England Journal of Medicine write that “over 400,000 people downloaded the app and enrolled in the study, not because of any health problem but because they were curious and wanted the reassurance of high-tech, zero-effort heart monitoring.”
In the real world there are many people at high risk for AF (and other medical problems as well) and many of these are underserved by the healthcare system, but 40-year-olds with smart watches are not among them. It shouldn’t need to be said but research and resources should go where they are most urgently needed, not to assuage the fears of the worried well who are willing to spend dollars on illusory or vanishingly small medical benefits
AF is not a problem that the vast majority of 40 year olds need to worry about. But invasion of privacy is a real problem for nearly everyone today, including, perhaps especially, 40-year-olds with smart watches. The editors discuss the issue of health privacy violations, which can occur “sometimes because of negligent security and sometimes because of deliberate and deceptive misuse of personal data.” They go on to write: “The uncomfortable fact is that our personal health data have considerable financial value to those who want to use them in the myriad marketplaces connected to our $3.7 trillion health economy.” Unfortunately, the authors of the Apple Heart Study paper don’t even mention, much less deal with, these privacy issues in their published paper.
So what then is the value or purpose of the study? The answer is clear— at least for Apple. Apple wants to sell more watches and phones, and it believes it can do so by promoting the potential health benefits of its products. Apple isn’t a healthcare company and has no particular interest in advancing medical research beyond this purpose. But Apple is the richest company on earth and it can easily afford to spend the tens of millions of dollars it must have cost to run this study. This cost is just a rounding error for the company.
This is why the study included such a large percentage of young people— a population with little or no relevance to AF. If the study had restricted enrollment to older participants it would have been far more useful from a medical and research perspective. But of course Apple wants to sell its products to everyone, and high risk elderly people constitute only a tiny proportion of their market.
The Apple Heart Study was performed by top researchers at Stanford University and other institutions. It’s not hard to imagine why academic researchers, who are always desperately seeking new sources of funding, might want to tap into a rich and easy target like Apple, which after all must seem like the motherlode. But we shouldn’t let the business interests of Apple and other technology companies distort the medical research agenda. Is this really how we want medical research to be conducted?